Steps toward artificial intelligence
Computers & thought
Representations for space planning
Communications of the ACM
Hi-index | 0.00 |
Given the 5-10 year goal of a computer program to design buildings, this paper discusses a proposed outline of problem areas. These problems are considered stubborn but tractable and need be systematically resolved. A program is described which investigates the problem of combining wall-units into a wall which satisfies relation value specification between rooms. Architectural design automation has apparently progressed beyond location programs to the point at which individual researchers are investigating problem areas whose output may be a partial program rather than a complete package. The evaluation of such a project should then fall upon (1) the recognition of relevant problem areas and (2) its progress demonstrated by the program within the context of an overall, long term goal.